In-line inspection using the Magnetic Flux Leakage (MFL) technique is sensitive both to pipe wall geometry and pipe wall stresses. Therefore, MFL inspection tools have the potential to locate and characterize mechanical damage in pipelines. However, the combined influence of stress and geometry make MFL signals from dents and gouges difficult to interpret. Accurate magnetic models that can incorporate both stress and geometry effects are essential to improve the current understanding of MFL signals from mechanical damage. MFL signals from dents include a geometry component in addition to a component due to residual stresses. If gouging is present, then there may also be an additional magnetic contribution from the heavily worked material at the gouge surface. The relative contribution of each of these components to the MFL signal depends on the size and shape of the dent in addition to other effects such as metal loss, wall thinning, corrosion, etc.
Magnetic Finite Element Analysis (FEA) can be applied to model MFL signals from mechanical damage defects having various sizes, shapes, and configurations. These models included geometry effects, contributions due to elastic strain (either residual strain or strain due to in-service loading), and also magnetic behavior changes due to severe deformation. The modeled results were then compared with experimental MFL signal measurements on dents and gouges produced in the laboratory as well under “field” conditions. Magnetic FEA models were produced of circular dents as well as dents elongated in the pipe axial and pipe hoop directions. Residual stress patterns were predicted in and around the dent using stress FEA modeling. The magnetic effects of these predicted residual stresses were incorporated into the magnetic FEA model by modifying the magnetic permeability in stressed regions in and around the dent. The modeled stress and geometry contributions to the MFL signal were examined separately, and also combined for comparison with experimental MFL results. Agreement between modeled and measured MFL signals was generally good, and the measured MFL signals were used to validate and refine the models.